Fix JRA-55 downloader

[LisaBock]

Description

The downloading works now without RDA login. And the link to the data changed.

• Closes #issue_number
• Link to documentation:

---

Checklist

It is the responsibility of the author to make sure the pull request is ready to review. The icons indicate whether the item will be subject to the 🛠 Technical or 🧪 Scientific review.

• 🛠 This pull request has a descriptive title
• 🛠 Code is written according to the code quality guidelines
• 🛠 Documentation is available
• 🛠 Tests run successfully
• 🛠 The list of authors is up to date
• 🛠 Any changed dependencies have been added or removed correctly
• 🛠 All checks below this pull request were successful

New or updated data reformatting script

• 🛠 Documentation is available
• 🛠 The dataset has been added to the CMOR check recipe
• 🛠 The dataset has been added to the shared data pools of DKRZ and Jasmin by the @ESMValGroup/OBS-maintainers team
• 🧪 Numbers and units of the data look physically meaningful

---

[jlenh]

I think there might be changes coming as they are migrating their website today (cf. https://rda.ucar.edu/news/rda-gdex-relaunch/). I will have a look tomorrow!

[jlenh]

Thanks Lisa, the new link works perfectly!

I think there are few things to remove still from the old code that are not relevant anymore:

• Removing the temporary files is not needed anymore
  

  ESMValTool/esmvaltool/cmorizers/data/downloaders/datasets/jra_55.py

  Lines 123 to 126 in c31a664

  	if os.path.exists("Authentication.log"):
  	os.remove("Authentication.log")
  	if os.path.exists("auth.rda_ucar_edu"):
  	os.remove("auth.rda_ucar_edu")

• The use of the download_options parameter is not needed anymore
  

  ESMValTool/esmvaltool/cmorizers/data/downloaders/datasets/jra_55.py

  Lines 110 to 112 in c31a664

  	downloader.download_file(
  	url + f"/{channel}/{year}/" + fname, download_options
  	)

[LisaBock]

@jlenh I still have problems with the downloaded data, as the formatter is not working with them. The files are too small. Did you experience the same issue?

[jlenh]

@jlenh I still have problems with the downloaded data, as the formatter is not working with them. The files are too small. Did you experience the same issue?

The formatter raises PermissionError for some of the variables for me on Levante but I still get cmorised outputs. I will try to run it fully in my own data directory to avoid permission issues w/ the data folders on Levante.

[valeriupredoi]

many thanks Lisa 🍻 Ah just noticed the network comms issues - Julien is on it, I see

[jlenh]

I had another look and I cannot open the newly downloaded GRIB files as I always get the following error:
"ERROR: unreadable message"
It indeed looks like the files on the new system are corrupted or incomplete in a way because their size is greatly reduced (a few kb compared to a few Mb).
Maybe we need to switch the downloader to access the files through DIAS at https://search.diasjp.net/en/dataset/JRA55 instead?

On the other hand, I think I couldn't run the formatter on the already downloaded data on Levante because of the index files that are written when opening the GRIB files which cannot be done on my side as I do not have writing permission in the ESMValTool data folders.